diff mbox series

[v2] PM: sleep: wakeirq: fix wake irq warning in system suspend

Message ID 20240301092657.15528-1-qingliang.li@mediatek.com (mailing list archive)
State In Next
Delegated to: Rafael Wysocki
Headers show
Series [v2] PM: sleep: wakeirq: fix wake irq warning in system suspend | expand

Commit Message

Qingliang Li March 1, 2024, 9:26 a.m. UTC
When driver uses pm_runtime_force_suspend() as the system suspend callback
function and registers the wake irq with reverse enable ordering, the wake
irq will be re-enabled when entering system suspend, triggering an
'Unbalanced enable for IRQ xxx' warning. In this scenario, the call
sequence during system suspend is as follows:
  suspend_devices_and_enter()
    -> dpm_suspend_start()
      -> dpm_run_callback()
        -> pm_runtime_force_suspend()
          -> dev_pm_enable_wake_irq_check()
          -> dev_pm_enable_wake_irq_complete()

    -> suspend_enter()
      -> dpm_suspend_noirq()
        -> device_wakeup_arm_wake_irqs()
          -> dev_pm_arm_wake_irq()

To fix this issue, complete the setting of WAKE_IRQ_DEDICATED_ENABLED flag
in dev_pm_enable_wake_irq_complete() to avoid redundant irq enablement.

Reviewed-by: Dhruva Gole <d-gole@ti.com>
Signed-off-by: Qingliang Li <qingliang.li@mediatek.com>
---
 drivers/base/power/wakeirq.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

Comments

Johan Hovold March 4, 2024, 3:19 p.m. UTC | #1
On Fri, Mar 01, 2024 at 05:26:57PM +0800, Qingliang Li wrote:
> When driver uses pm_runtime_force_suspend() as the system suspend callback
> function and registers the wake irq with reverse enable ordering, the wake
> irq will be re-enabled when entering system suspend, triggering an
> 'Unbalanced enable for IRQ xxx' warning. In this scenario, the call
> sequence during system suspend is as follows:
>   suspend_devices_and_enter()
>     -> dpm_suspend_start()
>       -> dpm_run_callback()
>         -> pm_runtime_force_suspend()
>           -> dev_pm_enable_wake_irq_check()
>           -> dev_pm_enable_wake_irq_complete()
> 
>     -> suspend_enter()
>       -> dpm_suspend_noirq()
>         -> device_wakeup_arm_wake_irqs()
>           -> dev_pm_arm_wake_irq()
> 
> To fix this issue, complete the setting of WAKE_IRQ_DEDICATED_ENABLED flag
> in dev_pm_enable_wake_irq_complete() to avoid redundant irq enablement.
> 
> Reviewed-by: Dhruva Gole <d-gole@ti.com>
> Signed-off-by: Qingliang Li <qingliang.li@mediatek.com>

Thanks for the fix. Looks correct to me:

Reviewed-by: Johan Hovold <johan+linaro@kernel.org>

I think you should add back the Fixes tag from v1 and CC stable as well:

Fixes: 8527beb12087 ("PM: sleep: wakeirq: fix wake irq arming")
Cc: stable@vger.kernel.org      # 5.16

Note that WAKE_IRQ_DEDICATED_REVERSE was added in 5.16 by commit
259714100d98 ("PM / wakeirq: support enabling wake-up irq after
runtime_suspend called") so no need to try to backport any further than
that.

Johan
Rafael J. Wysocki March 5, 2024, 11:41 a.m. UTC | #2
On Mon, Mar 4, 2024 at 4:19 PM Johan Hovold <johan@kernel.org> wrote:
>
> On Fri, Mar 01, 2024 at 05:26:57PM +0800, Qingliang Li wrote:
> > When driver uses pm_runtime_force_suspend() as the system suspend callback
> > function and registers the wake irq with reverse enable ordering, the wake
> > irq will be re-enabled when entering system suspend, triggering an
> > 'Unbalanced enable for IRQ xxx' warning. In this scenario, the call
> > sequence during system suspend is as follows:
> >   suspend_devices_and_enter()
> >     -> dpm_suspend_start()
> >       -> dpm_run_callback()
> >         -> pm_runtime_force_suspend()
> >           -> dev_pm_enable_wake_irq_check()
> >           -> dev_pm_enable_wake_irq_complete()
> >
> >     -> suspend_enter()
> >       -> dpm_suspend_noirq()
> >         -> device_wakeup_arm_wake_irqs()
> >           -> dev_pm_arm_wake_irq()
> >
> > To fix this issue, complete the setting of WAKE_IRQ_DEDICATED_ENABLED flag
> > in dev_pm_enable_wake_irq_complete() to avoid redundant irq enablement.
> >
> > Reviewed-by: Dhruva Gole <d-gole@ti.com>
> > Signed-off-by: Qingliang Li <qingliang.li@mediatek.com>
>
> Thanks for the fix. Looks correct to me:
>
> Reviewed-by: Johan Hovold <johan+linaro@kernel.org>
>
> I think you should add back the Fixes tag from v1 and CC stable as well:
>
> Fixes: 8527beb12087 ("PM: sleep: wakeirq: fix wake irq arming")
> Cc: stable@vger.kernel.org      # 5.16
>
> Note that WAKE_IRQ_DEDICATED_REVERSE was added in 5.16 by commit
> 259714100d98 ("PM / wakeirq: support enabling wake-up irq after
> runtime_suspend called") so no need to try to backport any further than
> that.

Applied as 6.9 material, tags added as suggested above.

Thanks!
diff mbox series

Patch

diff --git a/drivers/base/power/wakeirq.c b/drivers/base/power/wakeirq.c
index 42171f766dcb..5a5a9e978e85 100644
--- a/drivers/base/power/wakeirq.c
+++ b/drivers/base/power/wakeirq.c
@@ -313,8 +313,10 @@  void dev_pm_enable_wake_irq_complete(struct device *dev)
 		return;
 
 	if (wirq->status & WAKE_IRQ_DEDICATED_MANAGED &&
-	    wirq->status & WAKE_IRQ_DEDICATED_REVERSE)
+	    wirq->status & WAKE_IRQ_DEDICATED_REVERSE) {
 		enable_irq(wirq->irq);
+		wirq->status |= WAKE_IRQ_DEDICATED_ENABLED;
+	}
 }
 
 /**